[Mesa-dev] [PATCH] gallivm: fix gather implementation a bit
sroland at vmware.com
sroland at vmware.com
Thu Sep 7 16:16:23 UTC 2017
From: Roland Scheidegger <sroland at vmware.com>
gather is defined in terms of bilinear filtering, just without the filtering
part. However, there's actually some subtle differences required in our
implementation, because we use some tricks to simplify coord wrapping for the
two coords per direction.
For bilinear filtering, we don't care if we end up with an incorrect
texel, as long as the filter weight is 0.0 for it. Likewise, the order of
the texels doesn't actually matter (as long as they still have the correct
filter weight).
But for gather, these tricks lead to incrorrect results.
Fix this for CLAMP_TO_EDGE, and add some comments to the other wrap functions
which look broken (the 3 mirror_clamp plus mirror_repeat) (too complex to fix
right now, and noone really seems to care...).
---
src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c | 58 +++++++++++++++++++----
1 file changed, 48 insertions(+), 10 deletions(-)
diff --git a/src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c b/src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c
index cb4660e..64e08ef 100644
--- a/src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c
+++ b/src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c
@@ -299,6 +299,7 @@ lp_build_coord_repeat_npot_linear(struct lp_build_sample_context *bld,
*/
static void
lp_build_sample_wrap_linear(struct lp_build_sample_context *bld,
+ boolean is_gather,
LLVMValueRef coord,
LLVMValueRef length,
LLVMValueRef length_f,
@@ -388,13 +389,29 @@ lp_build_sample_wrap_linear(struct lp_build_sample_context *bld,
/* clamp to length max */
coord = lp_build_min_ext(coord_bld, coord, length_f,
GALLIVM_NAN_RETURN_OTHER_SECOND_NONNAN);
- /* subtract 0.5 */
- coord = lp_build_sub(coord_bld, coord, half);
- /* clamp to [0, length - 0.5] */
- coord = lp_build_max(coord_bld, coord, coord_bld->zero);
- /* convert to int, compute lerp weight */
- lp_build_ifloor_fract(&abs_coord_bld, coord, &coord0, &weight);
- coord1 = lp_build_add(int_coord_bld, coord0, int_coord_bld->one);
+ if (!is_gather) {
+ /* subtract 0.5 */
+ coord = lp_build_sub(coord_bld, coord, half);
+ /* clamp to [0, length - 0.5] */
+ coord = lp_build_max(coord_bld, coord, coord_bld->zero);
+ /* convert to int, compute lerp weight */
+ lp_build_ifloor_fract(&abs_coord_bld, coord, &coord0, &weight);
+ coord1 = lp_build_add(int_coord_bld, coord0, int_coord_bld->one);
+ } else {
+ /*
+ * The non-gather path will end up with coords 0, 1 if coord was
+ * smaller than 0.5 (with correpsonding weight 0.0 so it doesn't
+ * really matter what the second coord is). But for gather, we
+ * really need to end up with coords 0, 0.
+ */
+ coord = lp_build_max(coord_bld, coord, coord_bld->zero);
+ coord0 = lp_build_sub(coord_bld, coord, half);
+ coord1 = lp_build_add(coord_bld, coord, half);
+ /* Values range ([-0.5, length_f - 0.5], [0.5, length_f + 0.5] */
+ coord0 = lp_build_itrunc(coord_bld, coord0);
+ coord1 = lp_build_itrunc(coord_bld, coord1);
+ weight = coord_bld->undef;
+ }
/* coord1 = min(coord1, length-1) */
coord1 = lp_build_min(int_coord_bld, coord1, length_minus_one);
break;
@@ -424,6 +441,13 @@ lp_build_sample_wrap_linear(struct lp_build_sample_context *bld,
coord = lp_build_add(coord_bld, coord, offset);
}
/* compute mirror function */
+ /*
+ * XXX: This looks incorrect wrt gather. Due to wrap specification,
+ * it is possible the first coord ends up larger than the second one.
+ * However, with our simplifications the coordinates will be swapped
+ * in this case. (Albeit some other api tests don't like it even
+ * with this fixed...)
+ */
coord = lp_build_coord_mirror(bld, coord);
/* scale coord to length */
@@ -474,6 +498,20 @@ lp_build_sample_wrap_linear(struct lp_build_sample_context *bld,
offset = lp_build_int_to_float(coord_bld, offset);
coord = lp_build_add(coord_bld, coord, offset);
}
+ /*
+ * XXX: This looks incorrect wrt gather. Due to wrap specification,
+ * the first and second texel actually end up with "different order"
+ * for negative coords. For example, if the scaled coord would
+ * be -0.6, then the first coord should end up as 1
+ * (floor(-0.6 - 0.5) == -2, mirror makes that 1), the second as 0
+ * (floor(-0.6 - 0.5) + 1 == -1, mirror makes that 0).
+ * But with our simplifications the second coord will always be the
+ * larger one. The other two mirror_clamp modes have the same problem.
+ * Moreover, for coords close to zero we should end up with both
+ * coords being 0, but we will end up with coord1 being 1 instead
+ * (with bilinear filtering this is ok as the weight is 0.0) (this
+ * problem is specific to mirror_clamp_to_edge).
+ */
coord = lp_build_abs(coord_bld, coord);
/* clamp to length max */
@@ -929,7 +967,7 @@ lp_build_sample_image_linear(struct lp_build_sample_context *bld,
*/
if (!seamless_cube_filter) {
- lp_build_sample_wrap_linear(bld, coords[0], width_vec,
+ lp_build_sample_wrap_linear(bld, is_gather, coords[0], width_vec,
flt_width_vec, offsets[0],
bld->static_texture_state->pot_width,
bld->static_sampler_state->wrap_s,
@@ -940,7 +978,7 @@ lp_build_sample_image_linear(struct lp_build_sample_context *bld,
x11 = x01;
if (dims >= 2) {
- lp_build_sample_wrap_linear(bld, coords[1], height_vec,
+ lp_build_sample_wrap_linear(bld, is_gather, coords[1], height_vec,
flt_height_vec, offsets[1],
bld->static_texture_state->pot_height,
bld->static_sampler_state->wrap_t,
@@ -951,7 +989,7 @@ lp_build_sample_image_linear(struct lp_build_sample_context *bld,
y11 = y10;
if (dims == 3) {
- lp_build_sample_wrap_linear(bld, coords[2], depth_vec,
+ lp_build_sample_wrap_linear(bld, is_gather, coords[2], depth_vec,
flt_depth_vec, offsets[2],
bld->static_texture_state->pot_depth,
bld->static_sampler_state->wrap_r,
--
2.7.4
More information about the mesa-dev
mailing list